May 12 Dissertation Defense – Mari Miyao


PhD in Second Language Studies Dissertation Defense

Mari Miyao

The processing of referential expressions in discourse
by Chinese and Japanese L2 learners of English

Chair: Bonnie D. Schwartz

Friday, May 12, 2:00–4:00 p.m.
Moore Hall, Room 155A

Referential expressions (REs), such as proper names and pronouns, pose challenges to second language learners (L2ers). Generally, adult native speakers prefer a more explicit RE form (e.g., a proper name) when referring to a less salient/accessible entity in the discourse and a reduced RE form (e.g., a pronoun) for a salient/accessible entity (e.g., Ariel, 1990). To use REs in a successful manner, L2ers need to calculate the accessibility of discourse entities and associate the accessibility level with particular RE forms. The present study asks how adult L2ers whose first language (L1) is Chinese or Japanese (null-subject languages) comprehend and produce REs in discourses in English (a non-null-subject language).

Experiment 1 looks at the comprehension of REs in subject position (subject-REs), using closely-translated versions of a sentence-by-sentence self-paced reading task adapted from Gordon, Grosz, and Gilliom (1993), by native speakers of English, Chinese, and Japanese. Experiment 2 examines the same participants’ production of subject-REs via a three-panel picture-narration task adapted from Arnold and Griffin (2007). Experiments 3 and 4 employ the English version of the two tasks to explore the comprehension and production of subject-REs by low-intermediate-to-advanced L1-Chinese and L1-Japanese L2ers of English.

In the reading task, native English speakers and native Japanese speakers preferred the most reduced subject-REs—respectively, pronominal subjects and null-pronoun subjects—for accessible entities; native Japanese speakers strongly dispreferred overt-pronoun subjects. Different reading-time analyses indicated different subject-RE preferences for native Chinese speakers, but they never dispreferred overt-pronoun subjects. In the production task, all three language groups exhibited similar subject-RE preferences for accessible entities, but when the accessibility was reduced by the presence of another entity in the discourse or by a shifted discourse focus, they produced more explicit subject-REs (i.e., repeated names) more frequently.

Neither L2er group showed subject-RE preferences in reading, but in production, where pictures helped build firm discourse representations, they clearly preferred pronominal subjects for accessible entities and repeated-name subjects for entities with reduced accessibility. Overall, the present study suggests that when sufficient contextual support is provided, L2ers can calculate discourse accessibility and choose subject-RE forms according to the accessibility level (contra Sorace, 2011).